Segmentation on EndoVis 18
43.9mIoUMultiMAE
Evaluation results
| Method | Links | |
|---|---|---|
| MultiMAEEvaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=Surgical, Input Modality=RGB-D2026.01 | 43.9 | |
| Mask3DEvaluation Protocol=End-to-End Fine-tuning, Input Modality=Point Cloud2026.01 | 40.6 | |
| MultiMAE (IN-1k)Evaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=ImageNet-1k, Input Modality=RGB2026.01 | 39.4 | |
| DINOv2-RGBDEvaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=Surgical, Input Modality=RGB-D2026.01 | 26.9 | |
| DINOv2-RGBEvaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=Surgical, Input Modality=RGB2026.01 | 26.2 | |
| DINOv2 (LVD)Evaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=LVD, Input Modality=RGB2026.01 | 25.8 | |
| MAEEvaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=Surgical, Input Modality=RGB2026.01 | 23.5 | |
| MAE (IN-1k)Evaluation Protocol=End-to-End Fine-tuning, Pre-training dataset=ImageNet-1k, Input Modality=RGB2026.01 | 21.1 |